Audio and video transcription online

Free mode is unlimited transcription with no card and no subscription. Run an interview, several short calls, or a batch of voice messages — no limits.

On clean Russian or English speech the average error rate is about 8% — typical for a strong neural speech model. The transcript is ready for light editing, not a rewrite.

Yes. Speech recognition is trained on real-life speech, so it handles Russian regional accents, varied English pronunciation and code-switching. A heavy accent adds some errors but the text stays readable.

Yes, it automatically splits the conversation by voice and labels turns as "Speaker 1", "Speaker 2". You rename them in one click and the change applies across the whole transcript.

It confidently separates up to 10 voices on a single recording — enough for meetings, focus groups and podcasts. On large panels you can fine-tune speaker names manually in the editor.

Yes. Recordings from the street, an office room, or noisy Zoom calls still come out readable. Heavy noise adds some errors but context and speakers are preserved.

Yes, as long as the whisper is audible without heavy volume drops. On very quiet fragments the service may mark a gap — boosting the volume and re-uploading usually fixes it.

Yes. Phone recordings, IP-telephony call captures, and recordings from a bad connection are handled correctly, with slightly more errors than studio audio.

Russian, English, Spanish, German, French, Italian, Chinese, Japanese, Arabic and dozens more. Free speech recognition works the same on every language — no extra setup needed.

Yes. When meetings switch between languages or use English terms, the engine understands both and assembles a single transcript without losses.

The base vocabulary covers most professional terms in RU and EN. For rare words and abbreviations there is a glossary — add your own terms and the engine spells them correctly across transcripts.

Any typo is fixed directly in the editor — the change is saved with that transcript. In the glossary you can permanently pin team names, products and jargon so future transcripts spell them right automatically.

MP3, WAV, M4A, OGG, OPUS, FLAC, AAC, AMR, WMA and voice notes from messengers. MP3 transcription and voice recorder files all go through the same upload.

Yes. MP3 is the most common voice-recorder and podcast format — the service converts MP3 to text with timestamps and speakers in 5–10 minutes per hour of audio.

Yes, a WAV file is uploaded the same way as MP3 and often comes out slightly more accurate thanks to uncompressed quality. Great for studio recordings, podcasts and microphone captures.

Yes. iPhone Voice Memos are saved as M4A and can be uploaded directly — no separate MP3 conversion. It is the fastest way to turn an iPhone recording into text.

Yes. Telegram, WhatsApp and other messengers save voice notes as OGG/OPUS — the service recognises them directly and returns text within seconds.

Yes, the service extracts the audio track from MP4, MOV, MKV, WebM and AVI and runs video-to-text online. No separate conversion is required.

Yes. Save the WhatsApp voice note to a file or forward it to the bot — converting a voice message to text takes seconds and the result is saved to your account.

Yes. Telegram voice notes work via direct file upload. Voice messages, video notes and audio files are turned into text within seconds.

Up to 4 GB per file — roughly 8 hours of audio or a full HD webinar. Long recordings are automatically split into chunks and merged back into one document.

Up to 8 hours per file, which covers almost any meeting, lecture or interview. For longer recordings, split into parts and upload them sequentially — the texts can be merged via folders.

Yes, files are queued and processed in parallel. On the Team plan you can drop dozens of recordings at once — the service balances the load and preserves order.

Yes. A long lecture, interview or meeting is uploaded as one file — the service keeps context and timestamps across the full length. Speaker labels stay consistent from start to end.

Upload the WAV as-is — no re-encoding. Uncompressed audio gives slightly higher accuracy than MP3 or OGG. After transcription you can export to DOCX, SRT, TXT, JSON or Markdown.

Standard Android recorders save to M4A, AMR or 3GP — all three are accepted directly. Even with a non-standard codec, the service still extracts the audio track.

Yes. Paste the video URL — the service downloads the track itself and transcribes it. This is how YouTube, RuTube, VK Video and other platforms work without manual downloading.

Yes. Paste the link — the service downloads the audio track and produces a YouTube video-to-text transcript with timestamps and speakers. Great for notes, summaries and subtitles.

Yes. RuTube video-to-text works the same way as YouTube — just paste the link. Useful for creators publishing on RU platforms who need quick transcripts.

Yes. VK Video to text and Dzen recordings work by link with no manual download. The service grabs the audio and immediately runs transcription with speaker separation.

Yes. Download the Zoom recording (MP4 or M4A) and upload it — the service transcribes the meeting and assembles a meeting protocol with tasks and decisions.

Yes. Google Meet, Microsoft Teams and other conference recordings are uploaded as a file or by link. The output is a transcript with speakers, a summary and an action item list.

Yes. Transcribing Yandex Telemost, MTS Link and other Russian conferencing tools is a core scenario. Upload the recording file or link and you immediately get the meeting protocol.

Yes. Recordings from Skype, Discord or TeamSpeak are uploaded directly and the service separates participants by voice. With many speakers, the names are easy to set with a click in the editor.

Download the file from Google Drive or Yandex Disk and upload it — or paste a direct media link and the service fetches it for you. For large files, a direct download link is the most convenient option.

No need to download manually: the service downloads the track from the link itself. If the file is already on your disk, just drag it into the window — the transcript and timestamps will be the same.

An hour of audio is processed in roughly 5–10 minutes — dozens of times faster than a freelance transcriber. Short voice notes are returned within 5–20 seconds.

Typically 5–10 minutes per hour of recording. That is dozens of times faster than manual transcription, which takes 5–8 hours per hour of audio — and without any back-and-forth.

For now the service works with finished files and links — this gives the highest accuracy. Live dictation and streaming transcription are on the roadmap.

Files go into a queue and are processed in parallel — users do not block each other. Pro and Team plans get higher queue priority than free uploads.

Large files are processed fast — even a multi-hour recording is ready much quicker than its length. There is nothing to configure: you upload the file and get the finished text.

A 30–60 second voice note is transcribed in 5–10 seconds. Great for handling a stream of Telegram or WhatsApp voice notes during the workday.

Yes. Every turn is labelled with a speaker number, names are edited in one click and applied across the whole text. Especially useful for meetings, interviews and panel discussions.

Yes. Click a paragraph and the player jumps to that second, making it easy to verify the transcript against the original. Timestamps are exported with the text into JSON and SRT.

Yes. SRT subtitles are exported with one click — ready for YouTube, VK Video, RuTube and any video editor. Online SRT subtitles can be translated into other languages right in your account.

Yes, WebVTT is generated alongside SRT. This is what web players, HLS streams and social media plugins need — you get online video subtitles in both formats at once.

The service runs transcription first and then assembles the subtitles automatically from the timestamps. You do not have to read the text — just hit export to download SRT and VTT.

Yes. After transcription a single click produces a summary: key points, topics and decisions. You can generate a neural video summary in a minute and share it right away.

Yes. The service automatically prepares an online meeting protocol: key points, decisions, and an action list with owners and deadlines. The structure is close to a classic meeting protocol — templates and examples are available in-app.

Yes. The report shows a dedicated block of who promised what and by when — the action list can be exported straight into Jira, Trello, Notion or your team task tracker.

DOCX, PDF, TXT, Markdown, JSON and SRT/VTT subtitles — all in one click from the editor. You can export clean text or a version with timestamps and speakers.

Yes. DOCX opens in Word, Pages and Yandex Documents while preserving styles and speakers. PDF is convenient for client delivery or approvals — clean paginated layout, no reflow.

Yes. A finished transcript can be translated into English, Spanish, German, French, Chinese and other languages — the original stays alongside for reference. Useful for international teams and podcasts.

Yes. The built-in chat answers from the recording content: "what was said at minute 25", "what objections came up", "who owns what". A faster way to find specific moments than scrolling the text.

Yes. A unified search finds phrases across every saved transcript and highlights context with timestamps. Useful for journalists, researchers and product managers with dozens of interviews.

Yes. The glossary pins terms, employee names, products and acronyms — the service applies them automatically without typos. This significantly improves accuracy on niche topics.

Yes. The editor uses click-to-seek: tap a word and the player jumps to that second. This makes it easy to verify tricky parts and edit the transcript without manual scrubbing.

Yes. Revisions are saved automatically — you can roll back to any previous version and see what changed. Useful for collaborative text work.

The original audio is kept for 45 days, the playback copy for 210, the transcript text for 180 and edit snapshots for 7. You can delete a recording and its text sooner from your account, and deletion is final.

You and the people you have given access to.

Files and transcripts are stored on servers in Russia. Traffic goes over an encrypted channel.

Data is stored within the Russian Federation. All transfer is encrypted.

Audio is stored for 45 days, the playback copy for 210 days, transcripts for 180 days and editor snapshots for 7 days. After that, data is deleted automatically. A delete button is next to every transcript and the file is wiped from storage with no recovery option.

Yes. Only you can access the file, transfer is encrypted (HTTPS/TLS). Audio and transcript confidentiality is enforced at the infrastructure level.

Only people you invite, with configurable roles (viewer, editor, admin). The access log is visible and invitations can be revoked in one click.

Yes, this is a core scenario. The service transcribes Yandex Telemost, Zoom, Google Meet, MTS Link and other conferencing tools, producing a meeting protocol, action items and decisions.

Yes. Interview-to-text is a flagship scenario: the service automatically splits questions and answers by speaker and adds timestamps. Quotes can be highlighted in the editor and exported as a batch.

Yes. Podcast transcripts come with speakers, chapters and topic tags — handy for show notes and announcements. The text exports to Markdown or DOCX for CMS and distribution platforms.

Yes. Lecture transcription is popular among students and teachers: you get the full text plus a summary that is convenient for revision. Timestamps make it easy to jump back to any fragment.

Yes. Call transcription and call-to-text work with standard IP-telephony and CRM formats (MTS, MANGO, UIS, Bitrix). The output can be summarised by objections, goals and rep commitments.

Yes. In legal work exact quotes with timestamps are critical, and the service provides both. Data is stored within the Russian Federation, owner-only access.

Yes. Online interview transcription, focus groups and in-depth research are exactly what the service was built for. You can search across all interviews, tag quotes and assemble a thematic report.

Yes, subject to patient consent and local regulations. Data is stored within the Russian Federation, audio can be deleted right after text is ready, and medical terms can be pinned in the glossary in advance.

Registration takes about 10 seconds: email or social login. Without an account we cannot save your transcript and keep it visible only to you.

Everything runs in the browser — Chrome, Safari, Firefox, Edge on any OS. No separate app required, and the web version is mobile-friendly for phone uploads.

Yes. The transcription API lets you upload files and fetch transcripts and summaries from your own systems. Docs are in the help section, API key auth, JSON responses.

Yes. When a transcript is ready the service posts a webhook to your URL with metadata, so you can trigger client delivery, CRM save or editor import immediately.

Via API and webhooks you can pull finished transcripts into your own service. Direct integrations with Notion, Bitrix24 and amoCRM are planned.

Russian is recognised more accurately, the UI is in Russian, payments work with Russian cards and data is stored within the Russian Federation. Critical for organisations with data-security requirements.

Built-in Zoom and Telemost transcription returns raw text without an editor, search, DOCX/SRT export or task-aware summary. The service produces a full meeting document, not just a stream of words.

Many extensions run on third-party clouds and ship your file out unpredictably. The service processes recordings and stores data within the Russian Federation, making it suitable for confidential work recordings.

Yes — the service builds meeting minutes from an audio or video recording: agenda, key points, decisions, and action items with owners. Export to DOCX or PDF and circulate without rewriting from scratch.

The template includes date, participants, agenda items, summary of discussion, decisions, and action items. It mirrors a classic operations or executive meeting protocol and is accepted in most organizations without edits.

Yes. For standups, weekly syncs, or executive meetings the service splits speakers, captures action items with deadlines and owners, and delivers the minutes right after the call instead of two days later.

The service highlights phrases with explicit decisions, deadlines, and owners — "Ivan will prepare the report by Friday" becomes a single action item line. Edit the list and export action items separately as CSV or push to your task tracker.

The service is built for transcribing finished recordings, not for live real-time dictation. For long dictated text, record it on your phone's voice recorder and upload the file — you'll get cleaned-up text instead of patching up errors from the system dictation.

Yes — recordings from any voice recorder turn into text: M4A from iPhone, AMR/3GP from Android, WAV and MP3 from external recorders. Works for long lecture or interview recordings too.

System dictation mixes up speakers, drops endings on long sentences, and won't process existing files. The service processes the whole recording, separates speakers, adds punctuation, and gives you editable text instead of a raw stream of words.

Transcription is the conversion of spoken audio or video into written text — word by word. The service runs automatic transcription online with no manual typing, freelance marketplaces, or hours fiddling with the player.

Upload a video file (MP4, MOV, MKV, WebM) or paste a YouTube/RuTube/VK Video link — the service extracts audio, transcribes speech, and returns text with timecodes and speaker labels. Large videos are processed in chunks, nothing needs to be downloaded locally.

Yes. An English recording is transcribed into English text, and a Russian translation can be requested in the transcript chat — ask for a fragment or for the whole conversation.

Paste a YouTube link in another language — the service transcribes the original speech and produces a Russian translation of the text plus SRT/VTT subtitles. Handy for foreign lectures, podcasts, and interviews without official Russian subs.

Fireflies and Otter target English and Google Meet/Zoom meetings, don't accept RU cards, and store data abroad. The service natively transcribes Russian with punctuation, accepts any file format or link, and stores data within the Russian Federation.

Yes. Paste the meeting link — the bot joins, records the call, and sends back a transcript with speakers, timestamps, and ready-made minutes. We support Yandex Telemost, MTS Link, SaluteJazz (SberJazz), and Kontur Talk, and the minutes surface decisions and tasks right away.

Yes. Send the bot a voice message, audio, video, or a meeting link right in the chat — and get a transcript, a summary, and an action items list. The bot transcribes voice messages and turns voice into text, and it also supports export and chat over the transcript.